I might be able to give some insight. I work in clinical trial management, particularly for human subject safety oversight, so regulatory stuff is something I can contribute here.
Joel is correct, FDA approval is required for this device to be considered a medical device. Luckily, FDA approval is easier to obtain for a medical device than a new pharmaceutical. Medical devices are approved via the Premarket Approval (PMA) application process. A single study confirming the devices capabilities and safety is normally enough to get approval. In contrast, pharmaceuticals are approved through the New Drug Application (NDA) process and drug development is characterized by Phases I through IV clinical trials. NDA's can sometimes take 10 years before a drug can be brought to market, and that's IF the drug gets past all Phases. Furthermore, the entire process can cost a company $800 million+.
The FDA classified medical devices in three categories. Class I devices are low risk and subject to the least regulatory burdens. An example of this is dental floss (yes, really). Class II devices are deemed a bit more risky and have to deal with more regulatory issues. This mostly entails proving the device’s safety and effectiveness. Condoms are classified as Class II devices. The last category is Class III. These are deemed the highest risk devices by the FDA and are subject to some hefty regulation. Class III devices have to be approved by FDA before they are marketed. Replacement heart valves are classified as Class III devices.
Since the Ada hand isn't integrated directly with the body via graft or electrode placed in tissue (i think), it'll likely fall under Class II. However, the whole Open Source thing is likely going to be an issue, as my government policy sadly has yet to catch up with technology. New things scare them all.
Let's assume though it's Class II. This means it'd need a premarket notification submission, or a 510(k). This means you need to take a few extra steps before being approved as a medical device. You need to have performance testing data that's normally confirmatory, 10-15% of devices often require clinical data (doubt this hand will though), and some require animal testing and other trials to confirm safety and efficacy. If clinical studies ARE required, you have to deal with feasibility and pivotal analysis, and possibly bench-animal-human data collection.
There's pretty detailed and comprehensive guidelines for getting this approved and the stages of review that occur. If you want more information, let me know!